Naturally-occurring populations of microorganisms exhibit a wide array of biochemical and metabolic diversity. Due in part to difficulties in isolating and culturing many microorganisms, a vast number of potentially valuable proteins and polypeptides present in these populations have escaped identification. Indeed, it has been estimated that less than one percent of the world's microorganisms have been cultured to date. There remains a pressing need for new approaches to the characterization of proteins, polypeptides and metabolites from as-yet uncultivated, unidentified microorganisms, and also from known microorganisms. (The term “protein” as used hereinafter should be understood to encompass peptides and polypeptides as well.) There also remains a need for new approaches to the identification and isolation of the genes encoding these proteins, so as to enable the modification and/or production of the proteins.
One approach to this problem has been described by Short in U.S. Pat. Nos. 5,958,672; 6,001,574, 6,030,779, and 6,057,103 (the contents of which are incorporated herein by reference). In this approach, a genomic DNA library is prepared directly from an environmental sample (e.g. a soil sample), with or without making an attempt to isolate or culture any organisms that might be present. The DNA library is expressed in E. coli, and the expressed proteins are screened for a property or activity of interest. Short alludes to, but does not describe or enable, the use of fungal host cells in this method.
The approach as described suffers from several serious disadvantages, one of which is that E. Coli does not effectively express genes having introns. Roughly 90% of the species of microorganisms in soil are eukaryotes (principally fungi), which generally do have introns in their genomic DNA. Given that there are already about 100,000 species of eumycotan fungi known, with an estimated 1,000,000 yet to be discovered (B. Kendrick, The Fifth Kingdom, Mycologue Publications 1999), the potential for protein and metabolite diversity is far higher among the fungal genomes, but the presence of introns puts most of the fungal protein and metabolite repertoire out of the reach of bacterial expression systems. Not only are many classes of enzymes (e.g., secretory fungal lignin peroxidases and manganese-dependent peroxidases) unique to fungi, but there are many fungal proteins, including enzymes (e.g. lignin peroxidases, A. niger invertase), that are glycosylated, and such proteins would not be glycosylated if expressed by E. coli. The much higher number and greater size and complexity of fungal genomes, the uniqueness of many fungal proteins, and the glycosylation of many fungal proteins, all indicate that the fraction of microbial protein and metabolite diversity in a given environmental sample that could be actually detected by bacterial expression of genomic DNA is considerably less than 10%.
Due in part to the spread of AIDS and the rising population of organ transplant recipients, there is a growing population of immune-compromised or immuno-supressed individuals, and the number and variety of fungal infections has grown apace (Infect. Med. 16:380–382, 385–386 (1999)). There is a need to identify and characterize proteins from pathogenic fungi in the ongoing search for new targets for anti-fungal drugs, which requires the capability to screen DNA libraries derived from fungal genomes. Again, the presence of introns in fungal genomes makes expression of genomic DNA libraries difficult in most currently available bacterial hosts. There has also been a rise in the prevalence of antibiotic-resistant bacterial infections, creating a need for high-throughput screening for new fungal metabolites having antibiotic activity.
Eukaryotic genomes of higher organisms are also too complex for comprehensive expression of DNA libraries in bacteria. When all eukaryotic species are considered, bacteria represent only about 0.3% of all known species (E. O. Wilson, “The Current State of Biological Diversity”, in Biodiversity, National Academy Press, Washington D.C., 1988, Chapter 1); thus the fraction of the world's genetic diversity accessible to bacterial expression systems is extremely limited.
To avoid problems with introns, it is possible to prepare a cDNA library and express it in bacteria. However, this approach relies upon the presence of RNA transcripts, and any genes not actively being transcribed will not be represented in the library. Many desirable proteins are expressed only under specific conditions (e.g., virulence factors in pathogenic fungi) and these conditions may not exist at the time the mRNA is harvested. Furthermore, in order to obtain sufficient RNA to prepare a cDNA library, it is necessary to culture a fair amount of the organism. For organisms in environmental samples that do not grow well in culture, or novel microorganisms for which appropriate culture conditions are unknown, sufficient RNA will not be readily or reliably obtained. In contrast, sufficient genomic DNA can be obtained from a very small number of individual cells by PCR amplification, using either random primers or primers designed to favor certain classes of genes. Finally, genes that are highly expressed in an organism will tend to be over-represented in the mRNA, and thus over-represented at the expense of minimally-expressed genes in a cDNA library. In order to have a high level of coverage of the mRNA species present, a much larger number of clones must be screened if a cDNA library is employed instead of a genomic library, since the latter will have a more nearly equal representation of the variety of genes present. Clearly it is more desirable to screen a genomic DNA library if at all possible.
Also, E. coli is incapable of secretion of many proteins, and thus is undesirable as a host cell for screening purposes where the screening relies upon secretion of the gene product. An additional disadvantage for E. coli, and for bacterial hosts in general, is that prokaryotes cannot provide many of the post-translational modifications required for the activity of numerous eukaryotic proteins. In addition to glycosylation, subunit cleavage, disulfide bond formation, and proper folding of proteins are examples of the post-translational processing often required to produce an active protein.
To ensure such processing one can sometimes use mammalian cells, but mammalian cells are difficult to maintain, require expensive media, and are not generally transformed with high efficiency. Such transformation systems are therefore not convenient for high-throughput screening of proteins, although efforts have been made to employ mammalian cells as hosts for cDNA library screening (Schouten et al., WO 99/64582). An approach involving fusion of transformed protoplasts with mammalian cells prior to library screening has been described (U.S. Pat. No. 5,989,814), but expression of the protein library occurs in bacteria or yeast prior to cell fusion. There have been efforts to modify glycosylation patterns enzymatically after expression in host cells (Meynial-Salles and Combes, J. Biotechnol., 46:1–14 (1996)), but such methods must be tailored for specific products and are not suitable for expression of proteins from a DNA library. More recently, Maras et al., Eur. J. Biochem., 249:701–707 (1997) (see also U.S. Pat. No. 5,834,251) have described a strain of Trichoderma reesei engineered to express human GlcNAc transferase I. The enzyme transfers N-acetylglucosainine to mannose residues on other expressed exogenous proteins, a first step toward more closely approximating natural mammalian products.
The use of yeast as host cells solves some of the above problems, but introduces others. Yeast tend to hyper-glycosylate exogenous proteins (Bretthauer and Castellino, 1999, Biotechnol. Appl. Biochem. 30:193–200), and the altered glycosylation patterns often render expressed mammalian proteins highly antigenic (C. Ballou, in Molecular Biology of the Yeast Sacccharomyces, J. Strathern et al., eds., Cold Spring Harbor Laboratory Press, NY, 1982, 335–360). Although yeast are capable of coping with a limited number of introns, they are not generally capable of handling complex genes from higher species such as vertebrates. Even genes from filamentous fungi are usually too complex for yeast to transcribe efficiently, and this problem is compounded by differences in expression and splicing sequences between yeast and filamentous fungi (see e.g., M. Innis et al., Science 1985 228:21–26). Despite these drawbacks, transformation and expression systems for yeast have been extensively developed, generally for use with cDNA libraries. Yeast expression systems have been developed which are used to screen for naturally secreted and membrane proteins of mammalian origin (Klein, et al, Proc. Natl. Acad. Sci. USA 1996 93:7108–7113; Treco, U.S. Pat. No. 5,783,385), and for heterologous fungal proteins (Dalboge and Heldt-Hansen, Mol. Gen. Genet. 243:253–260 (1994)) and mammalian proteins (Tekamp-Olson and Meryweather, U.S. Pat. No. 6,017,731).
The term “yeast” as used in the context of yeast expression systems generally refers to organisms of the order Saccharomycetales, such as S. cerevisiae and Pichia pastoris. For the purposes of this disclosure, the terms “fungi” and “fungal” should be understood to refer to Basidiomycetes, Zygomycetes, Oomycetes, and Chythridiomycetes, and Ascomycetes of the class Euascomycetes, which are not of the order Saccharomycetales. Filamentous fungi may be distinguished from yeast by their hyphal elongation during vegetative growth, and obligately aerobic carbon catabolism (vegetative growth in yeast is accomplished by budding from a unicellular thallus, and yeast may employ fermentative catabolism.)
Proper intron splicing, and glycosylation, folding, and other post-translational modifications of fungal gene products would be most efficiently handled by a fungal host species, making filamentous fungi superior hosts for screening genomic DNA from soil samples. It also makes them excellent hosts for the production of fungal enzymes of commercial interest, such as proteases, cellulases, and amylases. It has also been found that filamentous fungi are capable of transcribing, translating, processing, and secreting the products of other eukaryotic genes, including mammalian genes. The latter property makes filamentous fungi attractive hosts for the production of proteins of biomedical interest. Glycosylation patterns introduced by filamentous fungi more closely resemble those of mammalian proteins than do the patterns introduced by yeast. For these reasons, a great deal of effort has been expended on the development of fungal host systems for expression of heterologous proteins, and a number of fungal expression systems have been developed. For reviews of work in this area, see Maras et al., Glycoconjugate J, 16:99–107 (1999); Peberdy, Acta Microbiol. Immunol. Hung. 46:165–174 (1999); Kruszewsa, Acta Biochim. Pol. 46:181–195 (1999); Archer et al., Crit. Rev. Biotechnol. 17:273–306 (1997); and Jeenes et al., Biotech. Genet. Eng. Rev. 9:327–367 (1991).
High-throughput expression and assaying of DNA libraries derived from fungal genomes would also be of use in assigning functions to the many mammalian genes that are currently of unknown function. For example, once a fungal protein having a property of activity of interest is identified, the sequence of the encoding gene may be compared to the human genome sequence to look for homologous genes.
Yelton et al., U.S. Pat. No. 4,816,405, discloses the modification of filamentous Ascomycetes to produce and secrete heterologous proteins. Buxton et al., in U.S. Pat. No. 4,885,249, and in Buxton and Radford, Mol. Gen. Genet. 196:339–344 (1984), discloses the transformation of Aspergillus niger by a DNA vector that contains a selectable marker capable of being incorporated into the host cells. McKnight et al., U.S. Pat. No. 4,935,349, and Boel, in U.S. Pat. No. 5,536,661, disclose methods for expressing eukaryotic genes in Aspergillus involving promoters capable of directing the expression of heterologous genes in Aspergillus and other filamentous fungi. Royer et al., in U.S. Pat. No. 5,837,847, and Berka et al., in WO 00/56900, disclose expression systems for use in Fusarium venenatum employing natural and mutant Fusarium spp. promoters. Conneely et al., in U.S. Pat. No. 5,955,316, disclose plasmid constructs suitable for the expression and production of lactoferrin in Aspergillus. Cladosporium glucose oxidase had been expressed in Aspergillus (U.S. Pat. No. 5,879,921).
Similar techniques have been used in Neurospora. Lambowitz, in U.S. Pat. No. 4,486,533, discloses an autonomously replicating DNA vector for filamentous fungi and its use for the introduction and expression of heterologous genes in Neurospora. Stuart et al. describe co-transformation of Neurospora crassa spheroplasts with mammalian genes and endogenous transcriptional regulatory elements in U.S. Pat. No. 5,695,965, and an improved strain of Neurospora having reduced levels of extracellular protease in U.S. Pat. No. 5,776,730. Vectors for transformation of Neurospora are disclosed in U.S. Pat. No. 5,834,191. Takagi et al. describe a transformation system for Rhizopus in U.S. Pat. No. 5,436,158. Sisniega-Barroso et al. describe a transformation system for filamentous fungi in WO 99/51756, which employs promoters of the glutamate dehydrogenase genes from Aspergillus awamori. Dantas-Barbosa et a., FEMS Microbiol. Lett. 1998 169:185–190, describe transformation of Humicola grisea var. thernoidea to hygromycin B resistance, using either the lithium acetate method or electroporation.
Among the more successful fungal expression systems are those of Aspergillus and Trichoderma, for example as disclosed by Berka et al. in U.S. Pat. No. 5,578,463; see also Devchand and Gwynne, J. Biotechnol. 17:3–9 (1991) and Gouka et al., Appl. Microbiol. Biotechnol. 47:1–11 (1997). Examples of transformed strains of Myceliophthora thermophila, Acremonium alabamense, Thielavia terrestris and Sporotrichum cellulophilum are presented in WO 96/02563 and U.S. Pat. Nos. 5,602,004, 5,604,129 and 5,695,985, which describe certain drawbacks of the Aspergillus and Trichoderma systems and suggest that other fungi may be more suited to large scale protein production.
Methods for the transformation of phyla other than Ascomycetes are known in the art; see for example Munoz-Rivas et al., Mol. Gen. Genet. 1986 205:103–106 (Schizophyllum commune); van de Rhee et al., Mol. Gen. Genet. 1996 250:252–258 (Agaricus bisporus); Arnau et al., Mol. Gen. Genet. 1991 225:193–198 (Mucor circinelloides); Liou et al., Biosci. Biotechnol. Biochem. 1992 56:1503–1504 (Rhizopus niveus); Judelson et al., Mol. Plant Microbe Interact. 1991 4:602–607 (Phytophthora infestans); and de Groot et al., Nature Biotechnol. 1998 16:839–842 (Agaricus bisporus).
In addition to the usual methods of transformation of filamentous fungi, such as for example protoplast fusion, Chakraborty and Kapoor, Nucleic Acids Res. 18:6737 (1990) describe the transformation of filamentous fungi by electroporation. De Groot et al., in Nature Biotechnol. 16: 839–842 (1998), describe Agrobacterium tumefaciens-mediated transformation of several filamentous fungi. Biolistic introduction of DNA into fungi has been carried out; see for example Christiansen et al., Curr. Genet. 29:100–102 (1995); Durand et al., Curr. Genet. 31:158–161 (1997); and Barcellos et al., Can. J. Microbiol. 44:1137–1141 (1998). The use of magnetic particles for “magneto-biolistic” transfection of cells is described in U.S. Pat. Nos. 5,516,670 and 5,753,477, and is expected to be applicable to filamentous fungi.
It is evident that much work has been done to develop expression systems using fungi as hosts. However, the common fin gal hosts are all filamentous fungi, which tend to form entangled mats of mycelia in unstirred cultures, and highly viscous suspension (submerged) cultures in stirred tank bioreactors. These properties of filamentous fungi also cause some problems in the industrial production of enzymes in fungal host cells. For example, high viscosity and/or the local formation of dense aggregates of mycelium, leads to difficulties in agitation, aeration, and nutrient diffusion. In general, filamentous fungi are not amenable to micropipetting of suspension cultures into microtiter plates, due to the viscosity of the cultures. Furthermore, due to the entangled mycelia, a culture of a typical filamentous fungus expressing a DNA library is not easily separated into separate clones on a large scale, which prevents evaluation of the individual genotypes as would be required in a high-throughput assay system.
Typical filamentous fungi, in the absence of constant agitation, tend to grow in the form of mats on the surface of a liquid culture medium, where they produce aerial spores. They do not generally sporulate when in submerged culture. Both of these properties present substantial obstacles to the culture of filamentous fungal clones in mircotiter plates, and to the efficient manipulation and use of such cultures for high-throughput screening. Suspended spores or other reproductively competent elements would suitable for separation and distribution into individual microtiter wells, whereas the production of aerial spores will lead to cross-contamination of microtiter wells if surface mats are allowed to form. Agitation of the medium in microtiter wells, to the extent needed to prevent mat formation, is not feasible. In addition to the problem of difficult-to-control aerial spores, surface mats interfere with light transmission, making many assays (in particular spectrophotometric absorbance assays) diffcult or impossible. Surface mats also interfere with processes such as oxygenation, reagent and nutrient addition, and pipetting.
The influence of fungal morphology on the physical properties of the culture has been recognized, and naturally-occurring strains having more favorable morphology have been identified, as described for example by Jensen and Boominathan in U.S. Pat. No. 5,695,985. Homogeneous distribution of loose mycelium, with pronounced branching, was described as a particularly desirable morphology. Schuster and Royer, in international patent application WO 97/26330 and U.S. Pat. No. 6,184,026, suggest a similar method of identifying fungal cells having more suitable morphology for industrial production of heterologous proteins. The method comprises screening mutants of a parent fungal cell line, rather than wild-type strains, to find a specific altered morphology, transforming the mutant, and assessing whether a culture of the transformed mutant produces more heterologous protein than the parent cell line. Mutants with at least 10% greater hyphal branching are particulary claimed. The method is illustrated for strains of Trichoderma, Fusarium and Aspergillus, and is suggested to be applicable to numerous other genera.
The effect of branching frequency on culture viscosity of Aspergillus oryzae mutants was examined by Bocking et al., Biotechnol. Bioeng. 65:638–648 (1999); more highly branched strains exhibited lower viscosity in this study. Van Wezel et al, in PCT application WO 00/00613, describe methods for reducing the branching and/or enhancing the fragmentation of filamentous microorganisms, whereby the viscosity of the culture is reduced. The method involves transforming the microorganisms with the SsgA gene of Streptomyces griseus. The method is demonstrated in filamentous bacteria of the order Actinomycetales, but is stated to be applicable to filamentous fungi. Dunn-Coleman et al., in WO 00/56893, describe an HbrA2 mutant A. nidulans, which exhibits a hyperbranched phenotype when grown above 42 ° C., and noted a linear relationship between the degree of hyphal branching and culture viscosity.
Most prior efforts in the field of filamentous fungal expression systems have been directed to the identification of strains suitable for industrial production of enzymes, and therefore attention has been focused on culture viscosity, stability of transformation, yield of heterologous protein per unit volume, and yield as a percentage of biomass. DNA libraries have been expressed in fungi; see for example Gems and Clutterbuck, Curr. Genet. 1993 24:520–524, where an Aspergillus nidulans library was expressed in A nidulans and Gems et al., Mol. Gen. Genet. 1994 242:467–471 where a genomic library from Penicillium was expressed in Aspergillus. Neither of these reports disclosed or suggested screening the expressed proteins; it was through complementation of mutant alleles in the host that the expression of genes from the DNA library was demonstrated. The complementation method requires a specific mutant host for each exogenous protein activity one wishes to detect, and does not provide a tool for general library screening.
The cloning of an Aspergillus niger invertase gene by expression in Trichoderma reesei was described by Berges et al., Curr. Genet. 1993 24:53–59. Using an A. niger genomic library constructed in a cosmid vector containing a selectable marker, and using as the host T. reesei (which is incapable of utilizing sucrose), an A. niger invertase gene was cloned by a sib selection procedure. Here, again, a very specific characteristic of the host was required to detect the presence of a single expressed exogenous protein, and screening of the genomic library was not disclosed or enabled.
The characteristics of a fugal host cell suitable for expression of a DNA library are different in many respects from the characteristics of hosts suitable for industrial protein manufacture. In general terms, a suitable fugal host for high-throughput screening should meet numerous criteria; among them are the following:                The host must be transformed with high efficiency.        The host must process intron-containing genes and carry out any necessary splicing.        The host must post-translationally process the expressed protein so that it is produced in an active form.        Where the library is to be assayed for a protein, the host must produce the protein in high enough yield for detection by the assay.        The host should accept a variety of expression regulatory elements, for ease of use and versatility.        The host should permit the use of easily-selectable markers.        The host cell cultures should be of low viscosity.        The host should be deficient in proteases and/or be anemable to suppression of protease expression.        The host must permit screens for a wide variety of exogenous protein activities or properties.        The hyphae in a culture of the host fungus should not be so entangled as to prevent the isolation of single clones, and should not be so entangled as to raise the viscosity to the point of preventing efficient transfer and replication in a miniaturized high throughput screening format (e.g. by micropipeting).        The host should not form surface mats, but should preferentially grow as a submerged culture.        The host should allow the efficient production of submerged spores or other propagules under the growth conditions provided in the high throughput screen.        
In cases where metabolites are being screened for, it would be advantageous if the host cells secreted the metabolites into the medium, where they could be readily detected and/or assayed. Ideally, the host should secrete only the exogenous protein.
In cases where a protein is being assayed for, it would be particularly advantageous if the host also expressed enough heterologous protein to enable isolation and purification of the protein. A host cell with this characteristic would make it possible to further characterize all heterologous proteins of interest merely by culturing the host cells, without the time-consuming molecular biological manipulations neeed to transfer the gene to another organism. Preferably, the host should be capable of secretion of the protein, as this would permit more reliable and more varied assays.
It would also be advantageous if the host cell were amenable to ready isolation of the heterologous DNA, so that further studies and modifications of the gene itself may be carried out.
In addition to these qualities of the host, the transformation system should also exhibit certrain characteristics. The transformation frequency should be sufficiently high to generate the numbers of transformants required for meaningful screens. Ideally, expression of the exogenous protein will be induced by a single inducer, by a single pathway, acting on a single promoter.
To date, no combination of host cells and transformation system has been developed that meets all, or even most, of these criteria. A need therefore remains for fungal host cell and transformation systems that are capable of efficiently expressing the gene products of a DNA library, especially genomic and/or eukaryotic genomic DNA libraries.